Rigetti Explained: Chiplets and the 108-Qubit Cepheus Machine

Updated 2026-10-11 | 4 min read | QUANTUM (QNT) community

The chiplet idea in plain English

Making one huge, perfect chip is hard, because every extra qubit adds another chance for a manufacturing flaw. Classical chip makers solved a similar problem with chiplets: build small pieces, test them, keep the good ones, and join them into a bigger package. Rigetti applies the same logic to quantum. The Cepheus-1-108Q is reported to be made of twelve interconnected 9-qubit chiplets, which gives 108 qubits. If you want the physics of the qubits themselves, start with superconducting qubits explained.

The timeline as reported

• January 2026: Rigetti said general availability of the 108-qubit system would slip to around the end of the first quarter of 2026. The company cited complexities with tunable couplers (the parts that switch interactions between qubits on and off) and said it would run another chip iteration. At that point it reported 99% median two-qubit fidelity on the 108-qubit system versus a 99.5% target, plus 99.7% on a 9-qubit system and 99.6% on a 36-qubit system.
• April 7, 2026: Rigetti announced general availability of Cepheus-1-108Q, reported through Rigetti Quantum Cloud Services and Amazon Braket. At launch it reported 99.1% median two-qubit gate fidelity and 99.9% median single-qubit fidelity.
• Q2 2026 earnings call: Rigetti's CEO was reported to expect fidelity to improve on the 108-qubit system before the end of 2026, while testing upgrades carefully before pushing them to cloud users.

The longer term goal reported for Rigetti is roughly 1,000 qubits at 99.9% two-qubit fidelity with gate speeds under 40 nanoseconds in about three years. That is a target, not a result, and I did not find a formal updated roadmap published after those reports.

What 99.1% actually means

Fidelity is the chance a gate does what you asked. A 99.1% two-qubit gate fails about once in every hundred tries. That sounds small, but a useful algorithm may need thousands or millions of gates in a row, so errors pile up fast. This is why error correction exists: it combines many noisy physical qubits into fewer, more reliable logical ones. Error correction schemes such as the surface code only work well when physical gates are already good enough, and the gap between 99.1% and 99.9% matters a great deal. The jump from 99% to 99.9% cuts errors by roughly a factor of ten.

Why chiplets are a real strength

The January 2026 numbers show something useful. The 9-qubit and 36-qubit systems reached 99.7% and 99.6%, while the 108-qubit system trailed. That says the building block is strong and the hard part is wiring up the whole package: the couplers, the links between chiplets and the calibration of hundreds of settings. Chiplet designs let a company improve one small piece and reuse it everywhere, so progress can compound. If Rigetti closes the gap, the same recipe can scale to larger systems.

What to watch next

• Whether the median two-qubit fidelity on the 108-qubit machine reaches the 99.5% target on cloud systems that customers can use, not just in lab runs.
• Whether Rigetti publishes the formal roadmap update it was reported to plan for the second half of 2026.
• Independent benchmarks. See quantum benchmarks explained for why a single fidelity number is only part of the story.

An honest and optimistic read

Delays are not failures. Complex hardware almost always runs late, and Rigetti was open about the reason and shipped within about three months of the announced slip. Reaching general availability of a 108-qubit machine on public clouds is a real step: researchers and developers can now run circuits on it. The remaining task, getting error rates down, is the same task every hardware maker faces.

Rigetti is a publicly listed company, but this guide is not an assessment of its stock or any asset, and the QNT memecoin has no connection to Rigetti or to any other company. Do your own research and never treat any article as financial advice.

Sources and further reading

• The Quantum Insider: Rigetti updates timeline for Cepheus-1-108Q (January 2026)
• Quantum Computing Report: Rigetti releases Cepheus-1-108Q (April 2026)
• Nasdaq: Rigetti Q2 earnings call highlights (call summary of management guidance, including the 99.5% target)

Reported as of 2026-10-09. Company roadmaps are targets and often slip. Check each company's own announcements. Nothing here is financial advice, and the QNT memecoin is independent of Quantinuum Ltd and of every company named.

Frequently asked questions

Share on X

Keep reading

• Superconducting Qubit Makers Compared: Rigetti, IQM, Alice & Bob, OQC, Fujitsu and Anyon
  A plain English side by side look at six superconducting quantum hardware makers, what each is betting on, and the 2025 to 2026 milestones that have been reported.
• Quantum Benchmarks Explained: Why Qubit Count Misleads
  How do you measure a quantum computer? Learn what error rates, fidelity, coherence time and quantum volume mean, and why qubit count alone is not enough.
• Quantum Error Correction Explained
  Qubits are fragile, so quantum computers need error correction. Learn how logical qubits are built and why this is the key challenge.
• Dilution Refrigerators, Wiring and Chip Fabrication: The Hardware Behind Superconducting Qubits
  Why superconducting quantum chips live near absolute zero, how a dilution refrigerator gets there stage by stage, and why wiring and fabrication are the real scaling limits.

All Quantum industry, people and AI guides | Back to top | Search the site

Main pages: Quantum computing explained | Quantum and crypto | Companies | Quantum news | Glossary